- Proposal for an AI Liability Directive is completely inadequate for consumers. vzbv calls for no-fault liability for AI systems.
- In reality, the proposed directive would mean consumers have almost no chance of obtaining compensation.
- The burden of proof on consumers is unrealistically high.
The European Commission has published a proposal for a liability directive specifically aimed at artificial intelligence. From a consumer perspective, we welcome the fact that the European Commission is taking steps to address the issue. Unfortunately, however, the current proposal is not sufficient to ensure adequate compensation for consumers who suffer harm.
The European Commission’s AILD adopts the approach of fault-based liability. This means consumers must prove the provider is at fault in order to obtain compensation. The burden of proof on consumers is so high that in practice it will be almost impossible for consumers to do so successfully. This is also because AI systems are notoriously lacking in transparency and are highly complex. The AILD’s disclosure requirement for providers against whom legal action is taken fails to remedy the problem.
The Federation of German Consumer Organisations (Verbraucherzentrale Bundesverband – vzbv) recommends the introduction of strict liability (no-fault liability). Damage caused to consumers when an AI system is used as intended should be sufficient for provider liability. Consumers would then have a realistic chance of actually obtaining compensation.